An image-assisted fluorescence spectroscopy detection method and system for additive manufacturing

Abstract

This invention proposes an image-assisted fluorescence spectroscopy detection method and system for additive manufacturing. The method includes the following steps: acquiring slice images of the additively manufactured parts to be detected and performing image preprocessing to obtain a standardized image containing several outlines of the parts to be detected; performing edge detection and outline recognition on the standardized image to extract the center coordinates and corresponding vertex coordinates of each outline in the image coordinate system and constructing an image coordinate set; obtaining the transformation ratio coefficient between the image coordinate system and the actual physical coordinate system through a multi-point calibration method, establishing a linear mapping relationship between the two, and converting the image coordinate set into the actual coordinate set; based on the center coordinates, vertex coordinates, and detection spot size of each outline in the actual coordinate set, generating a detection path covering all areas to be detected. This method realizes full-process automation and intelligence of elemental composition detection in additive manufacturing, significantly improving detection efficiency and reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of spectral detection technology, and in particular to an image-assisted fluorescence spectral detection method for additive manufacturing parts. BACKGROUND

[0002] With the rapid development of additive manufacturing technologies such as selective laser melting and electron beam melting, the manufacturing of complex structure and function integrated parts has become increasingly common. In order to ensure the structural reliability and functional stability, the material composition detection of additive manufacturing parts after production is particularly critical. Fluorescence spectrometer, as a commonly used element analysis tool, has the advantages of non-destructive, fast and high sensitivity, and is widely used in material characterization. However, there are still significant challenges in actual detection.

[0003] Publication No. CN106626377B discloses an additive manufacturing method and device for real-time detection of powder bed surface deformation, which includes controlling the scanning of the powder bed surface by a ray to form a grating line; controlling the imaging device to image the grating line and determining whether the grating line has deformation according to the imaging result; stopping the additive manufacturing when the grating line has deformation and the deformation amount is greater than the allowable value; adjusting the energy of the ray according to the deformation amount when the grating line has deformation and the deformation amount is less than or equal to the allowable value; adjusting the energy of the ray according to the deformation amount includes: when the deformation amount is positive, the energy of the ray is lowered; when the deformation amount is negative, the energy of the ray is increased; melting the powder bed surface by the ray; and melting the powder bed surface by the ray before and / or after controlling the scanning of the powder bed surface by the ray.

[0004] Although the above-mentioned prior art can realize online monitoring of powder bed deformation, it relies on a fixed mode of grating scanning path and lacks path planning mechanism, resulting in low detection efficiency and a large amount of redundant scanning. At the same time, this method does not involve the high-precision mapping relationship between image coordinates and physical coordinates, and cannot realize accurate positioning of the detection position, thereby reducing the accuracy and efficiency of spectral detection of additive manufacturing parts. SUMMARY

[0005] Therefore, the present application provides an image-assisted fluorescence spectral detection method and system for additive manufacturing parts, which significantly reduces the redundant scanning path by image coordinate conversion and combined with the honeycomb snake wing path planning algorithm, greatly improves the detection efficiency while ensuring full coverage and no blind area detection, thereby improving the accuracy and efficiency of spectral detection of additive manufacturing parts.

[0006] The technical solution of the present application is as follows: In a first aspect, the present application provides an image-assisted fluorescence spectral detection method for additive manufacturing, comprising the following steps:

[0007] S1, acquire a slice image of an additive manufacturing part to be detected, and perform image preprocessing to obtain a standardized image containing contours of a plurality of parts to be detected;

[0008] S2, using a pre-trained network model, edge detection and contour recognition are performed on the standardized image to extract the center coordinates and corresponding vertex coordinates of each part contour in the image coordinate system, and an image coordinate set is constructed;

[0009] S3, the conversion proportion coefficient between the image coordinate system and the actual physical coordinate system is obtained by a multi-point calibration method, a linear mapping relationship between the two is established, and the image coordinate set is converted into an actual coordinate set based on the linear mapping relationship;

[0010] S4, according to the center coordinates, vertex coordinates of each contour in the actual coordinate set, and the detection spot size, a honeycomb snake path arrangement algorithm is used to generate a detection path covering all the areas to be detected, and a scanning point coordinate file is output;

[0011] S5, control the motion system to read the scanning point coordinate file, drive the fluorescence spectrum detector to move to each scanning point according to the detection path, and trigger the spectrum collection in turn to complete the element composition detection.

[0012] On the basis of the above technical scheme, preferably, step S1 includes the following sub-steps:

[0013] Acquire a slice image of an additive manufacturing part to be detected;

[0014] Based on the HSV color space model, set the HSV threshold range corresponding to the boundary color; use a color segmentation algorithm to extract the positioning boundary region;

[0015] Crop the image in the positioning boundary region and adjust it to a predetermined standard size to obtain a standardized image containing contours of a plurality of parts to be detected.

[0016] On the basis of the above technical scheme, preferably, in step S3, the conversion proportion coefficient between the image coordinate system and the actual physical coordinate system is obtained by a multi-point calibration method, including the following sub-steps:

[0017] Select a plurality of distributed reference points on the slice image and obtain their coordinates in the image coordinate system;

[0018] Measure the actual coordinates of each reference point in the physical coordinate system; calculate the ratio of the corresponding distances of each reference point in the image coordinate system and the physical coordinate system respectively to obtain a plurality of proportion coefficients;

[0019] Linear regression fitting is performed on the plurality of proportion coefficients by using the least squares method to obtain the final conversion proportion coefficient between the image coordinate system and the actual physical coordinate system.

[0020] Preferably, in the step S3, the multiple distributed reference points on the slice image are selected, including:

[0021] A copper foil is arranged around the edge of the printed substrate as a boundary reference point for spatial positioning on the slice image;

[0022] The fluorescence spectrum detector is controlled to scan the spot on the printed substrate;

[0023] When the spot crosses the edge of the copper foil, the copper element signal strength shows a step mutation, and the first derivative peak of the copper element signal is taken as the physical boundary of the printed substrate.

[0024] Preferably, in the step S3, the linear mapping relationship between the two is established, and the image coordinate set is converted into the actual coordinate set based on the linear mapping relationship, including the following sub-steps: obtaining the actual coordinate corresponding to the center point of the printed substrate in the physical coordinate system; calculating the physical coordinate after the image coordinate conversion by multiplying the image coordinate and the final conversion scale factor and adding the actual coordinate corresponding to the center point of the printed substrate in the physical coordinate system; and constructing the actual coordinate set based on the converted actual physical coordinate.

[0025] Preferably, in the step S4, the following sub-steps are included:

[0026] The actual width and height of each part contour are calculated according to the vertex coordinates, and the actual width and height are compared with the spot size;

[0027] If the actual width and height are both smaller than the spot size, it is determined that the part contour is a small-size area, and the center point of the part contour is taken as the only scanning point;

[0028] If the actual width and height are greater than the spot size, it is determined that the part contour is a large-size area, and the X-direction scanning step is set to be equal to the spot diameter, and the Y-direction scanning step is set to be equal to a preset proportion multiple of the spot diameter;

[0029] Based on the center coordinates of the part contour, the spot diameter is dynamically adjusted according to the X-direction scanning step, the Y-direction scanning step, and the width and height of the part contour image, and the initial rectangular scanning point array covering the part contour is generated, with the constraint condition that the edges of the spot are tangent to the image boundaries;

[0030] The initial rectangular scanning point array is adjusted by staggering, the X coordinates of all scanning points in even rows or odd rows are adjusted to offset by a spot radius length in the X direction, the scanning points of adjacent rows are staggered in the X direction, and a honeycomb distribution is formed.

[0031] All the scanning points are connected in rows to form detection paths, and the connection directions of the paths of two adjacent rows are opposite, forming a snake-shaped reciprocating scanning path.

[0032] On the basis of the above technical scheme, preferably, the step S4 further comprises the following sub-steps:

[0033] The distance between the rightmost column of scanning points in the initial rectangular scanning point array and the right boundary of the contour is calculated, and if the distance is greater than 0, a column of scanning points is supplemented on the right boundary;

[0034] The distance between the lowermost row of scanning points in the initial rectangular scanning point array and the bottom boundary of the contour is calculated, and if the distance is greater than 0, a row of scanning points is supplemented on the lowermost boundary;

[0035] Finally, a detection path covering all the areas to be detected is generated.

[0036] On the basis of the above technical scheme, preferably, the step S5 comprises the following sub-steps:

[0037] By integrating the fluorescence spectrum detector into the control motion system, the control motion system is used to move the fluorescence spectrum detector in three-axis directions;

[0038] The control motion system reads the scanning point coordinate file, drives the fluorescence spectrum detector to move to each scanning point according to the detection path, and sends a trigger signal to the fluorescence spectrum detector after positioning;

[0039] After receiving the signal, the fluorescence spectrum detector performs an element composition detection task at the current position, returns a completion signal after detection, and the motion system drives the fluorescence spectrum detector to move to the next scanning point, and the cycle is executed until the detection of all scanning points is completed.

[0040] In a second aspect, the present application also provides an image-assisted fluorescence spectrum detection system for additive manufacturing, which is realized by using an image-assisted fluorescence spectrum detection method for additive manufacturing, and the system comprises:

[0041] A preprocessing module is used to acquire slice images of an additive manufacturing part to be detected, and perform image preprocessing to obtain standardized images containing contours of a plurality of parts to be detected;

[0042] An image processing module is used to perform edge detection and contour recognition on the standardized images by using a pre-trained network model, extract center coordinates and corresponding vertex coordinates of each part contour in an image coordinate system, and construct an image coordinate set;

[0043] A coordinate conversion module is used to acquire conversion proportion coefficients between the image coordinate system and an actual physical coordinate system by a multi-point calibration method, establish a linear mapping relationship between the two, and convert the image coordinate set into an actual coordinate set based on the linear mapping relationship;

[0044] a path planning module, configured to generate a detection path covering all the to-be-detected areas according to the center coordinates, vertex coordinates of each contour in the actual coordinate set and the detection spot size, and output a scanning point coordinate file by using a honeycomb serpentine path arrangement algorithm;

[0045] a motion control module, configured to control the motion system to read the scanning point coordinate file, drive the fluorescence spectrum detector to move to each scanning point according to the detection path, and trigger spectrum collection in sequence to complete element composition detection.

[0046] In a third aspect, the present application further provides a computer-readable storage medium, which stores a program of an image-assisted fluorescence spectrum detection method for additive manufacturing, and the program of the image-assisted fluorescence spectrum detection method for additive manufacturing is executed to implement the image-assisted fluorescence spectrum detection method for additive manufacturing.

[0047] The image-assisted fluorescence spectrum detection method for additive manufacturing and the system thereof have the following beneficial effects over the prior art:

[0048] (1) The automatic positioning of the detection area is realized through image intelligent recognition and accurate mapping of physical coordinates, the optimized detection path covering all areas without blind spots is generated by combining the honeycomb serpentine path planning algorithm, and finally the detection efficiency, accuracy and reliability are significantly improved through the closed-loop linkage of the motion control system and the spectrum detector;

[0049] (2) The high-precision and high-robustness linear mapping relationship between the image space and the physical space is established by combining the multi-point calibration and the least squares fitting algorithm and the spectrum signal accurate positioning technology of the copper foil boundary, the coordinate deviation problem caused by the system cumulative error and mechanical backlash is solved, and the accuracy and repeatability of the entire detection system are significantly improved;

[0050] (3) The global optimization and full-coverage generation of the detection path are realized through the size adaptive judgment mechanism and the honeycomb serpentine path planning algorithm, not only the two-dimensional detection blind area is effectively eliminated by the honeycomb misplacement point arrangement, but also the idle time of the motion system is greatly reduced by using the serpentine reciprocating connection, and the detection is ensured without omission by using the supplement point strategy at the boundary, so that the integrity and reliability of the element composition analysis results are guaranteed while the detection efficiency is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1 Flow chart of the image-assisted fluorescence spectrum detection method for additive manufacturing of the present application;

[0053] Figure 2 Slice image of the image-assisted fluorescence spectrum detection method for additive manufacturing of the present application;

[0054] Figure 3 Structure schematic diagram of a three-axis motion system of the image-assisted fluorescence spectrum detection method for additive manufacturing of the present application;

[0055] Figure 4 Principle diagram of the image-assisted fluorescence spectrum detection system for additive manufacturing of the present application. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0057] As shown in Figures 1-3 the first aspect, the image-assisted fluorescence spectrum detection method for additive manufacturing of the present application comprises the following steps:

[0058] S1, acquiring a slice image of an additive manufacturing part to be detected, and performing image preprocessing to obtain a standardized image containing contours of a plurality of parts to be detected.

[0059] It should be noted that the embodiment step S1 first performs image acquisition and preprocessing steps, which aims to process the original slice image that may contain redundant information into a standardized image containing only effective detection area and uniform size, laying a foundation for subsequent image recognition.

[0060] The step S1 includes the following sub-steps:

[0061] Acquiring a slice image of an additive manufacturing part to be detected;

[0062] It should be noted that the slice image is derived from the slicing software used by the additive manufacturing equipment. Before or after the printing task starts, the system exports the two-dimensional slice image of the current printing layer from the slicing software. The image clearly shows all the areas on the printing layer where the metal powder will be melted and shaped, as well as the unoccupied substrate area.

[0063] Based on the HSV color space model, the HSV threshold range corresponding to the positioning boundary color is set; and a color segmentation algorithm is used to extract the positioning boundary region.

[0064] It should be noted that the preprocessing module first converts the input slice image of the RGB color model to the HSV color space. The HSV space is closer to the human eye's perception of color than the RGB space, and can separate color information, saturation information, and brightness information. It has better robustness to light changes and is more conducive to color-based segmentation. The system presets the HSV threshold range corresponding to the positioning boundary color. For example, if the boundary is red, the hue (H) range can be set to (0, 10) and (170, 180), the saturation (S) range can be set to (100, 255), and the brightness (V) range can be set to (50, 255). These thresholds can be input as parameters to adapt to different color boundaries. A color segmentation algorithm is used to perform threshold processing on the HSV image to generate a binary mask, where the white area represents the recognized red boundary and the black area represents other parts. Morphological operations are then used to fill small holes, and the largest contour is extracted using a contour finding algorithm to obtain the desired positioning boundary.

[0065] The image in the positioning boundary region is cropped and adjusted to a predetermined standard size to obtain a standardized image containing the contours of several parts to be detected.

[0066] According to the extracted positioning boundary contour, the minimum circumscribed rectangle is obtained. The system uses this rectangle as a reference to crop the original slice image, completely removing all redundant information outside the image and the red border itself, and only retaining the core printing area image inside the border. This step ensures that subsequent processing only focuses on the effective part contour. The image obtained after cropping is uniformly scaled or interpolated to a predetermined standard size, eliminating image size differences caused by different slice software settings or export resolutions. This provides a computational basis for the fixed proportional conversion relationship between image coordinates and physical coordinates in subsequent steps, ensuring the consistency, repeatability, and accuracy of system processing.

[0067] It can be understood that through intelligent boundary recognition and image standardization processing based on the HSV color space, redundant information and size differences in the original slice image are effectively eliminated, and a standardized image with uniform size and only containing the effective part contour is automatically extracted, laying a solid foundation for subsequent recognition and positioning, and significantly improving the automation, accuracy, and consistency of system processing.

[0068] S2, using a pre-trained network model to perform edge detection and contour recognition on the standardized image, extracting the center coordinates and corresponding vertex coordinates of each part contour in the image coordinate system, and constructing an image coordinate set.

[0069] It should be noted that the edge detection and contour recognition are performed on the standardized image by using the pre-trained semantic segmentation network model. For a regular rectangular contour, four corner points are directly extracted. For an irregular polygonal contour, a polygon approximation algorithm is used to obtain a key vertex sequence representing the shape characteristics of the contour. These vertex coordinates define the shape and boundary of the part. The geometric center is calculated based on all the pixel points of the contour or the vertex coordinates, and the center coordinates will serve as the representative position of the part contour in the image coordinate system.

[0070] The center coordinates and corresponding vertex coordinates of all part contours extracted in the above steps are stored in elements, and all elements contained in each slice layer are stored as an image coordinate sub-set. The image coordinate sub-sets are saved according to the slice layer sequence to form an image coordinate set, which completely describes the geometric information of all target parts to be detected in the printed part by additive manufacturing.

[0071] It can be understood that by introducing the pre-trained semantic segmentation network model, high-precision and automatic recognition and feature extraction of part contours in the standardized image are realized, and the shortcomings of poor adaptability of traditional image processing algorithms to complex and fuzzy edges are effectively overcome. By calculating the geometric center and key vertex coordinates and constructing a structured image coordinate set, complete and accurate geometric information is provided for subsequent coordinate mapping and path planning, greatly improving the intelligent level and recognition reliability of the detection system.

[0072] S3, the conversion proportion coefficient between the image coordinate system and the actual physical coordinate system is obtained by a multi-point calibration method, a linear mapping relationship between the two is established, and the image coordinate set is converted into an actual coordinate set based on the linear mapping relationship.

[0073] The conversion proportion coefficient between the image coordinate system and the actual physical coordinate system obtained by the multi-point calibration method in step S3 includes the following sub-steps:

[0074] Select a plurality of distributed reference points on the slice image and obtain their coordinates in the image coordinate system;

[0075] It should be noted that, on the standardized slice image obtained by step S1, n (n≥3) uniformly distributed and easily identifiable reference points are manually or automatically selected, which should cover the four corners and the center area of the image to ensure the representativeness and global accuracy of the calibration; n (n≥3) uniformly distributed and easily identifiable reference points are manually or automatically selected on the standardized slice image obtained by step S1. These reference points should cover the four corners and the center area of the image as much as possible to ensure the representativeness and global accuracy of the calibration. The pixel coordinates of the n reference points in the image coordinate system are accurately obtained using image processing software or algorithms (such as click point selection or corner point detection); on the corresponding actual printed substrate, using measurement equipment such as optical measuring instrument, laser displacement sensor or encoder reading of high-precision numerical control platform, the coordinates (X_r_i, Y_r_i) of the points corresponding to the above n reference points in the actual physical coordinate system are measured.

[0076] The actual coordinates corresponding to each reference point in the physical coordinate system are measured; the ratio of the corresponding distances of each reference point in the image coordinate system and the physical coordinate system is calculated respectively to obtain a plurality of proportionality coefficients;

[0077] Wherein, including: calculating the Euclidean distance of each two reference points (i, j) in the image coordinate system and the actual coordinate system;

[0078] The distance calculation expression in the image coordinate system is:

[0079]

[0080] In the formula, (X_i_i, Y_i_i) represents the image coordinates of the i-th reference point; (X_i_j, Y_i_j) represents the image coordinates of the j-th reference point;

[0081] The distance calculation expression in the physical coordinate system is:

[0082]

[0083] In the formula, (X_r_i, Y_r_i) represents the actual coordinates of the i-th reference point; (X_r_j, Y_r_j) represents the actual coordinates of the j-th reference point;

[0084] The proportionality coefficient generated by the pair of reference points is calculated, and the expression is:

[0085] k_ij=d_r_ij / d_i_ij;

[0086] By combining different pairs of reference points, C(n, 2) proportionality coefficients k_ij can be obtained.

[0087] The least square method is used to linearly regress and fit multiple proportional coefficients, and a final conversion proportional coefficient between the image coordinate system and the actual physical coordinate system is solved.

[0088] Since the image coordinates and the actual physical coordinates are in a linear proportional relationship, a model is established ; The model has no intercept because the zero pixel distance should correspond to the zero physical distance.

[0089] The least square method is used to linearly regress and fit multiple proportional coefficients, and a final conversion proportional coefficient between the image coordinate system and the actual physical coordinate system is solved.

[0090]

[0091] In the formula, m is the number of all effective reference point pairs, m=C(n,2); is the image distance of the i-th point pair; is the actual distance of the i-th point pair.

[0092] This method can resist the interference of random measurement errors, and can especially reduce the overall deviation caused by inaccurate measurement of individual point pairs, so as to obtain a more robust and more accurate global proportional coefficient.

[0093] The linear mapping relationship between the two is established in step S3, and the image coordinate set is converted into the actual coordinate set based on the linear mapping relationship, including the following sub-steps: obtaining the actual coordinates corresponding to the center point of the printing substrate in the physical coordinate system; calculating the physical coordinates of the image coordinates after conversion by multiplying the image coordinates and the final conversion proportional coefficient, and adding the actual coordinates corresponding to the center point of the printing substrate in the physical coordinate system; constructing the actual coordinate set based on the converted actual physical coordinates, and the expression is:

[0094]

[0095] In the formula, (X_c, Y_c) is the reference point coordinate of the origin of the image coordinate system in the physical coordinate system, that is, the center point of the printing substrate, and (X_i, Y_i) is the image coordinate.

[0096] The mapping relationship is applied to batch calculate the image coordinate set constructed in step S2, and finally the actual physical coordinate set used to guide the movement of the fluorescence spectrometer is obtained; the accurate mapping from the virtual image space to the real physical world is realized, which lays a solid theoretical foundation for subsequent automatic path planning and accurate positioning detection.

[0097] The multiple distributed reference points on the slice image are selected in step S3, including:

[0098] The copper foil is arranged around the edge of the printing substrate as a boundary reference point for spatial positioning on the slice image;

[0099] It should be noted that, in order to realize high-precision coordinate mapping, a stable and accurate actual physical coordinate system is first established. In the embodiment, a foil strip with high-contrast elements, preferably copper foil, is arranged around the edge of the printing substrate as a spatial positioning reference. The copper foil strip provides a clear, known physical boundary reference.

[0100] The fluorescence spectrum detector is controlled to scan the light spot on the printing substrate.

[0101] It should be noted that the fluorescence spectrum detector is controlled to finely scan the light spot in the region of the printing substrate containing the copper foil edge, and the characteristic X-ray fluorescence intensity signal of the copper element is collected in real time. When the light spot crosses the boundary between the copper foil and the substrate, the copper element signal intensity will show a stepwise mutation. By taking the derivative of the signal curve, the peak point of the first derivative is the accurate physical position of the copper foil boundary.

[0102] When the light spot crosses the copper foil edge, the copper element signal intensity shows a stepwise mutation. The first derivative peak of the copper element signal is used as the physical boundary of the printing substrate.

[0103] It should be noted that by scanning and locating the boundary points of the two side edges of the substrate, and combining the actual design size of the substrate, the accurate coordinates (X_c, Y_c) of the geometric center point of the printing substrate in the physical space can be calculated, and this point is defined as the origin of the entire actual physical coordinate system. This method effectively calibrates the cumulative error and mechanical backlash of the motion system itself, ensuring the accuracy and reproducibility of the coordinate system reference.

[0104] In addition, in the reference point selection process, specific marker points on the copper foil boundary or feature points at known positions on the substrate can be preferred.

[0105] It should be noted that by using the multi-point calibration and least squares fitting algorithm, combined with the spectral signal accurate positioning technology of the copper foil boundary, a high-precision and high-robustness linear mapping relationship between the image space and the physical space is established, solving the problem of coordinate deviation caused by system cumulative error and mechanical backlash. This lays a solid technical foundation for realizing full-automatic and high-reliability detection from virtual image recognition to physical space accurate positioning, and significantly improves the precision and repeatability of the entire detection system.

[0106] S4, according to the center coordinates, vertex coordinates of each contour in the actual coordinate set, and the detection light spot size, a honeycomb snake path arrangement algorithm is used to generate a detection path covering all the regions to be detected, and a scanning point coordinate file is output.

[0107] It should be noted that the actual physical coordinate set obtained according to step S3 is used to generate a high-efficiency blind-free detection path based on the spot characteristics of the fluorescence spectrometer. The path planning method comprehensively considers the detection efficiency and coverage integrity, and adopts an optimized strategy combining honeycomb staggered distribution and snake-shaped reciprocating connection to achieve the optimization.

[0108] Step S4 includes the following sub-steps:

[0109] The actual width and height of each part contour are calculated according to the vertex coordinates, and the actual width and height are compared with the spot size;

[0110] If the actual width and height are both less than the spot size, it is determined that the part contour is a small-size area, and the center point of the part contour is taken as the only scanning point;

[0111] If the actual width and height are greater than the spot size, it is determined that the part contour is a large-size area, and the X-direction scanning step is set to be equal to the spot diameter, and the Y-direction scanning step is set to be equal to a preset proportion multiple of the spot diameter;

[0112] It should be noted that each part contour in the actual coordinate set is traversed, and the actual width W and height H of the contour are calculated according to the vertex coordinates of the contour;

[0113] If W≤spot diameter D and H≤spot diameter D, it is determined that the contour is a small-size area. A single spot can completely cover it, and the system takes the geometric center point of the contour as its only scanning point;

[0114] If W>spot diameter D or H>spot diameter D, it is determined that the contour is a large-size area, and the subsequent path arrangement sub-step is entered to generate a scanning point array;

[0115] Taking the center coordinates of the part contour as the reference, the spot diameter is dynamically adjusted according to the X-direction scanning step, the Y-direction scanning step, and the width and height of the part contour image, and an initial rectangular scanning point array covering the part contour is generated, with the constraints that the spot edges are tangent to the image boundaries;

[0116] It should be noted that the X-direction scanning step dx=D; the Y-direction scanning step dy=D*R; Wherein, R is a preset proportion coefficient, the value range is usually between 0.5 and 0.9, and the preferred value is 0.75. This design is to form a certain spot overlap in the Y direction to ensure coverage.

[0117] Taking the center coordinates of the contour as the reference, the required number of rows and columns are calculated according to the width W and height H of the contour, with the constraints that the spot edges of the first and last rows / columns are tangent to the contour boundaries, and an initial rectangular scanning point array covering the area is generated.

[0118] The initial rectangular scanning point array is adjusted by offsetting the X coordinates of all scanning points in even rows or odd rows by a spot radius length, so that the scanning points of adjacent rows are staggered in the X direction, forming a honeycomb distribution;

[0119] All scanning points are connected by rows to form detection paths, and the connection directions of adjacent two rows are opposite, forming a snake-shaped reciprocating scanning path.

[0120] It should be noted that, in order to improve the uniformity of two-dimensional plane coverage and eliminate detection blind area, the initial rectangular point array is optimized; the X coordinates of all scanning points in even rows or odd rows are adjusted to offset by dx / 2 in the X direction; after the adjustment, the scanning points between adjacent rows are staggered in the X direction, thereby forming a honeycomb distribution on the two-dimensional plane, making the coverage of the spot in the Y direction more continuous, and effectively avoiding the stripe-shaped uncovered area that may be caused by the traditional chessboard point distribution. In order to maximize the detection efficiency and reduce the air return time of the motion platform, a snake-shaped scanning mode is used to connect all scanning points to form a detection path.

[0121] It should be noted that, in order to ensure complete coverage of the edge area, the system performs full coverage verification after generating the basic path to ensure that the generated scanning path achieves complete coverage of the entire target area without blind area. Step S4 further includes the following sub-steps:

[0122] The distance between the rightmost column of scanning points in the initial rectangular scanning point array and the right boundary of the contour is calculated. If the distance is greater than 0, a column of scanning points is supplemented on the right boundary;

[0123] The distance between the lowermost row of scanning points in the initial rectangular scanning point array and the bottom boundary of the contour is calculated. If the distance is greater than 0, a row of scanning points is supplemented on the lowermost boundary;

[0124] Finally, a detection path covering all the areas to be detected is generated.

[0125] It can be understood that, through the size adaptive judgment mechanism and the honeycomb snake-shaped path planning algorithm, global optimization and full coverage generation of the detection path are achieved. Not only the two-dimensional detection blind area is effectively eliminated by the honeycomb offset point distribution, but also the air time of the motion system is greatly reduced by using the snake-shaped reciprocating connection, and the supplement point strategy is used at the boundary to ensure that there is no omission in detection, thereby significantly improving the detection efficiency while ensuring the integrity and reliability of the element composition analysis results.

[0126] S5, control the motion system to read the scanning point coordinate file, drive the fluorescence spectrum detector to move to each scanning point according to the detection path, and trigger the spectrum collection in turn to complete the element composition detection.

[0127] It should be noted that step S5 is the final execution stage of realizing full-process automation, which converts the digital detection path generated by the foregoing steps into actual physical movement and detection action of the fluorescence spectrometer, and finally completes the collection of element components through collaborative communication and linkage control between subsystems.

[0128] Step S5 includes the following sub-steps:

[0129] By integrating the fluorescence spectrum detector into the control motion system, the control motion system is used to move the fluorescence spectrum detector in three-axis directions;

[0130] Before starting detection, ensure that the three-axis motion system integrated with the fluorescence spectrum detector has completed the zero return or origin positioning operation, so that its current position is aligned with the physical coordinate system origin (X_c, Y_c), and the three-axis motion system is composed of an X-Y-Z slide driven by a high-precision stepper motor or a servo motor.

[0131] The control motion system reads the scan point coordinate file, drives the fluorescence spectrum detector to move to each scan point according to the detection path, and sends a trigger signal to the fluorescence spectrum detector after positioning;

[0132] The motion control system reads the scan point coordinate file from the specified storage path, the system parses the file, loads the physical coordinate sequence in it into the memory in order, and forms a detection point queue to be executed;

[0133] After receiving the signal, the fluorescence spectrum detector performs the element component detection task at the current position, returns a completion signal after detection, and the motion system drives the fluorescence spectrum detector to move to the next scan point, and the cycle is executed until the detection of all scan points is completed.

[0134] After the detection process is completed, the system automatically releases the pause state, resumes the printing program, and starts the construction work of the next layer. It can be understood that through the closed-loop linkage of the motion control system, the fluorescence spectrum detector and the central control module, seamless conversion and accurate execution of the digital detection path to physical detection action are realized, high-precision three-axis positioning and synchronous triggering mechanism are used to ensure automatic collection of element component data at each detection point, non-destructive, multi-point and full-automatic detection process between printing layers is realized, and the real-time and precision of process quality control are significantly improved.

[0135] As shown in Figure 4 , in a second aspect, the present application also provides an image-assisted fluorescence spectrum detection system for additive manufacturing, which is realized by using an image-assisted fluorescence spectrum detection method for additive manufacturing, and the system comprises:

[0136] a preprocessing module, configured to acquire slice images of an additive manufacturing part to be detected, and perform image preprocessing to obtain standardized images containing contours of a plurality of parts to be detected;

[0137] an image processing module, configured to perform edge detection and contour recognition on the standardized images by using a pre-trained network model, extract center coordinates and corresponding vertex coordinates of each part contour in an image coordinate system, and construct an image coordinate set;

[0138] a coordinate conversion module, configured to acquire conversion proportion coefficients between the image coordinate system and an actual physical coordinate system by a multi-point calibration method, establish a linear mapping relationship therebetween, and convert the image coordinate set into an actual coordinate set based on the linear mapping relationship;

[0139] a path planning module, configured to generate a detection path covering all areas to be detected by using a honeycomb serpentine path arrangement algorithm according to the center coordinates, vertex coordinates of each contour in the actual coordinate set, and a detection spot size, and output a scanning point coordinate file;

[0140] a motion control module, configured to control a motion system to read the scanning point coordinate file, drive a fluorescence spectrum detector to move to each scanning point according to the detection path, and trigger spectrum acquisition in sequence to complete element composition detection.

[0141] It should be noted that the system corresponds to the above-mentioned image-assisted fluorescence spectrum detection method for additive manufacturing, and all implementation manners in the above-mentioned method embodiments are applicable to the embodiments of the system and can achieve the same technical effects.

[0142] Those skilled in the art can appreciate that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0143] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system and modules can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0144] In the embodiments of the present application, it should be understood that the disclosed system and method can be implemented in other manners. For example, the embodiments of the apparatus described above are merely schematic, and the division of the units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0145] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0146] In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0147] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or partially contribute to the prior art, or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk, and various other media that can store program codes.

[0148] Moreover, it is pointed out that, in the system and method of the present application, obviously, the components or steps can be decomposed and / or recombined. These decompositions and / or recombination should be considered as equivalents of the present application. Also, the steps of performing the above series of processes can naturally be executed in time sequence according to the order of description, but do not necessarily have to be executed in time sequence. Some steps can be executed in parallel or independently of each other. It can be understood by those skilled in the art that all or any steps or components of the method and device of the present application can be implemented in hardware, firmware, software or a combination thereof in any computing device (including processors, storage media, etc.) or network of computing devices, using the basic programming skills of those skilled in the art upon reading the description of the present application.

[0149] Therefore, the object of the present application can also be achieved by running a program or a set of programs on any computing system. The computing system can be a commonly known general-purpose system. Therefore, the object of the present application can also be achieved only by providing a program product containing program code for implementing the method or device. That is, such a program product also constitutes the present application, and a storage medium storing such a program product also constitutes the present application. Obviously, the storage medium can be any commonly known storage medium or any storage medium developed in the future. It is also pointed out that, in the device and method of the present application, obviously, the components or steps can be decomposed and / or recombined. These decompositions and / or recombination should be considered as equivalents of the present application. Also, the steps of performing the above series of processes can naturally be executed in time sequence according to the order of description, but do not necessarily have to be executed in time sequence. Some steps can be executed in parallel or independently of each other.

[0150] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An image assisted fluorescence spectroscopy detection method for additive manufacturing, characterized in that, The method comprises the following steps: S1, acquiring a slice image of an additive manufacturing part to be detected, and performing image preprocessing to obtain a standardized image containing contours of a plurality of parts to be detected; S2, using a pre-trained network model to perform edge detection and contour recognition on the standardized image, extracting the center coordinates and corresponding vertex coordinates of each part contour in the image coordinate system, and constructing an image coordinate set; S3, obtaining a conversion proportionality coefficient between the image coordinate system and the actual physical coordinate system through a multi-point calibration method, establishing a linear mapping relationship between the two, and converting the image coordinate set into an actual coordinate set based on the linear mapping relationship; The following sub-steps are included: Selecting a plurality of distributed reference points on the slice image and obtaining their coordinates in the image coordinate system; the selection of a plurality of distributed reference points on the slice image includes: arranging copper foils around the edges of the printing substrate as boundary reference points for spatial positioning on the slice image; controlling the fluorescence spectrum detector to scan the light spot on the printing substrate; when the light spot crosses the edge of the copper foil, the copper element signal strength appears a stepwise mutation, and the first derivative peak value of the copper element signal is taken as the physical boundary of the printing substrate; Measuring the corresponding actual coordinates of each reference point in the physical coordinate system; calculating the ratio of the corresponding distances of each reference point in the image coordinate system and the physical coordinate system respectively, and obtaining a plurality of proportionality coefficients; Using the least squares method to linearly regress and fit the plurality of proportionality coefficients to obtain the final conversion proportionality coefficient between the image coordinate system and the actual physical coordinate system; S4, according to the center coordinates, vertex coordinates and detection spot size of each contour in the actual coordinate set, using a honeycomb snake path arrangement algorithm to generate a detection path covering all the areas to be detected, and outputting a scanning point coordinate file; S5, controlling the motion system to read the scanning point coordinate file, driving the fluorescence spectrum detector to move to each scanning point according to the detection path, and triggering the spectrum collection in turn to complete the element composition detection.

2. The image-assisted fluorescence spectroscopy detection method for additive manufacturing of claim 1, wherein, The following sub-steps are included in step S1: Acquiring a slice image of an additive manufacturing part to be detected; Based on the HSV color space model, set the HSV threshold range corresponding to the positioning boundary color; use the color segmentation algorithm to extract the positioning boundary region; Crop the image in the positioning boundary region and adjust it to the preset standard size to obtain a standardized image containing contours of a plurality of parts to be detected.

3. The image-assisted fluorescence spectroscopy detection method for additive manufacturing of claim 1, wherein: The following sub-steps are included in step S3 for establishing a linear mapping relationship between the two and converting the image coordinate set into an actual coordinate set based on the linear mapping relationship: obtaining the corresponding actual coordinates of the center point of the printing substrate in the physical coordinate system; calculating the physical coordinates of the image coordinates after conversion by multiplying the image coordinates and the final conversion proportionality coefficient, and adding them to the corresponding actual coordinates of the center point of the printing substrate in the physical coordinate system; and constructing an actual coordinate set based on the converted actual physical coordinates.

4. The image-assisted fluorescence spectroscopy detection method for additive manufacturing of claim 3, wherein, The following sub-steps are included in step S4: According to the vertex coordinates, calculate the actual width and height of each part contour, and compare the actual width and height with the spot size; If the actual width and height are both less than the spot size, it is determined that the part profile is a small size area, and the center point of the part profile is taken as the only scanning point; If the actual width and height are both greater than the spot size, it is determined that the part profile is a large size area, and the X-direction scanning step is set to be equal to the spot diameter, and the Y-direction scanning step is set to be equal to a preset multiple of the spot diameter; Taking the center coordinates of the part profile as the reference, the spot diameter is dynamically adjusted according to the X-direction scanning step, the Y-direction scanning step, and the width and height of the part profile image, so that the edges of the spot are tangent to the image boundaries as the constraint condition, and an initial rectangular scanning point array covering the part profile is generated; The initial rectangular scanning point array is adjusted by staggering, and the X coordinates of all scanning points in even or odd rows are adjusted to be offset by a spot radius length in the X direction, so that the scanning points of adjacent rows are staggered in the X direction to form a honeycomb distribution; All scanning points are connected by row to form a detection path, and the connection directions of adjacent two rows are opposite to form a snake-shaped reciprocating scanning path.

5. The image-assisted fluorescence spectroscopy detection method for additive manufacturing of claim 4, wherein, Step S4 further includes the following sub-steps: The distance between the rightmost column of scanning points in the initial rectangular scanning point array and the right boundary of the profile is calculated, and if the distance is greater than 0, a column of scanning points is supplemented on the right boundary; The distance between the lowermost row of scanning points in the initial rectangular scanning point array and the bottom boundary of the profile is calculated, and if the distance is greater than 0, a row of scanning points is supplemented on the lowermost boundary; Finally, a detection path covering all the areas to be detected is generated.

6. The image-assisted fluorescence spectroscopy detection method for additive manufacturing of claim 1, wherein, Step S5 includes the following sub-steps: The fluorescence spectrum detector is integrated on the control motion system, and the control motion system is used to move the fluorescence spectrum detector in three-axis directions; The control motion system reads the scanning point coordinate file, drives the fluorescence spectrum detector to move to each scanning point according to the detection path, and sends a trigger signal to the fluorescence spectrum detector after positioning; After receiving the signal, the fluorescence spectrum detector performs element composition detection on the current position, returns a completion signal after detection, and the motion system drives the fluorescence spectrum detector to move to the next scanning point, and the cycle is executed until the detection of all scanning points is completed.

7. An image-assisted fluorescence spectroscopy detection system for additive manufacturing, implemented using the image-assisted fluorescence spectroscopy detection method for additive manufacturing according to any one of claims 1 to 6, characterized in that, The system comprises: a preprocessing module configured to acquire slice images of an additive manufacturing part to be detected, and perform image preprocessing to obtain standardized images containing a plurality of part profiles to be detected; an image processing module configured to perform edge detection and profile recognition on the standardized images by using a pre-trained network model, extract center coordinates and corresponding vertex coordinates of each part profile in an image coordinate system, and construct an image coordinate set; a coordinate conversion module configured to acquire conversion proportion coefficients between the image coordinate system and an actual physical coordinate system by a multi-point calibration method, establish a linear mapping relationship between the two, and convert the image coordinate set into an actual coordinate set based on the linear mapping relationship; a path planning module configured to generate a detection path covering all areas to be detected by using a honeycomb snake path arrangement algorithm according to the center coordinates, vertex coordinates of each profile in the actual coordinate set, and the detection spot size, and output a scanning point coordinate file. The motion control module is used for controlling the motion system to read the scanning point coordinate file, driving the fluorescence spectrum detector to move to each scanning point according to a detection path, and triggering spectrum collection in sequence to complete element composition detection.

8. A computer-readable storage medium, characterized in that, The storage medium has stored thereon a program of an image-assisted fluorescence spectrum detection method for additive manufacturing, and the program of the image-assisted fluorescence spectrum detection method for additive manufacturing, when executed, implements the image-assisted fluorescence spectrum detection method for additive manufacturing according to any one of claims 1-6.

Images